Great post from Philip SU about the slippage going on in the Microsoft Vista development.
There are some interesting, if meaningless, stats in here. The average US software developer writes 6200 lines of code in a year. In 1999, the same developer wrote 9000 lines of code. The 2000 developers working on Vista averaged 1000 lines of code over its development lifetime. I don't know what the numbers are like here at SuperMegaCorp, but I suspect it's way way more than that. Our product is nowhere near the size of Vista, but we suffer some of the same pains that the Microsoft development group is suffering.
The Vista delay is not just due to complex software and layers of process. To be fair, when you've got 2000 developers you need some process. One of the bigger problems is estimation. Development estimates on Vista get pushed back until they fit the plan, or are just ignored. "Every once in a while, Truth still pipes up in meetings. When this happens, more often than not, Truth is simply bent over an authoritative knee and soundly spanked into silence."
Does anyone really believe this will work? I haven't read that article in the Harvard Business Review, but the "Head in the Sand" theory of software project management doesn't strike me as a sound one. I suppose the idea is that if every level keeps lying about the feasibility of a ship date, it's only the top level that will get the axe when the date is blown by months. Although, sometimes that doesn't even happen, when it obviously should. The finger of blame conveniently gets pointed elsewhere instead.
I can remember a conversation that went something like this:
PM: When can you get this done?
Dev: 18 months.
PM: Ack! How about 6 months?
Dev: *Cough* *Splutter* Maybe 12 months.
PM: Let's schedule it for 9 months.
Dev: Whatever...
Guess what? It was delivered 18 months later, buggy as all-get-out. Despite the legions of well-meaning developers who were thrown into the project, since as everyone knows, the more developers you throw at a sinking ship the faster it gets done.
But hey, I'm just a developer, and the Harvard Business Review doesn't say you should trust your developers. If they were any good, they'd be in management, right?
(
Software development)
(
Management)